Anal. Chem. 2002, 74, 5112-5120
In Situ Raman Spectroscopy Studies of Metal Ion Complexation by 8-Hydroxyquinoline Covalently Bound to Silica Surfaces Rory H. Uibel and Joel M. Harris*
Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112-0850
Raman spectroscopy is applied to an investigation of the interfacial chemistry of silica-immobilized 8-hydroxyquinoline (8HQ) for binding of metal ions over a wide range of solution conditions. Since the derivatized silica has a high specific binding capacity, the mass of silica equilibrated with solution needs to be small for studies of reactions with trace-level (µM) metal ions; otherwise, the solution volume required to reach equilibrium becomes excessive. To address this problem, a small-volume flow cell is designed for this work using a fiber-optic Raman probe inserted directly into the packed end of a microcolumn, allowing excitation and collection of Raman scattering from less than 10 mg of derivatized silica. This cell is attached to a flow system that allows control of solution conditions while the response of the 8HQ-silica material is acquired by continuous monitoring of Raman scattering from the sample. Raman spectra of the deprotonated, neutral, protonated, and copper-complexed forms of the ligand can be distinguished, allowing protontransfer and metal ion binding reactions of the ligand to be investigated. To account for the effects of changing surface potential on these reactions, ζ-potential measurements are made on the 8HQ-silica particles under the same solution conditions that are employed in the Raman scattering measurements. The observed pH dependence of metal ion binding was corrected for the effect of surface potential using the Boltzmann equation, and the resulting equilibrium constant for binding of Cu2+ was independent of metal ion concentration over a 100-fold range from 30 µM to 5 mM. Silica gels that are chemically modified with surface-immobilized reagents can be used for the cleanup of metals from waste streams, for separation of different metal ion species, or preconcentration of metal ions for trace-level detection. Recent progress has been made in synthesis of new ligands that can be bound onto silica supports.1-6 One of the most widely used silica(1) Bruening, R. L.; Tarbet, B. J.; Krakowiak, K. E.; Bruening, M. L.; Izatt R. M.; Bradshaw, J. S. Anal. Chem. 1991, 63, 1014-1017. (2) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L.; Chem. Rev. 1991, 91, 1721-1785. (3) Jezorek, J. R.; Tang, J.; Cook, W. L.; Obie, R.; Ji, D.; Rowe, J. M. Anal. Chim. Acta 1994, 290, 303-315. (4) Lessi, P.; Moreira, J. C.; Filho, N. L.; Campos, J. T. Anal. Chim. Acta 1996, 327, 183.
5112 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002
immobilized metal-chelating reagents used for preconcentration of trace elements is 8-hydroxyquinoline (8HQ).7-10 This ligand exhibits a high affinity for a variety of transition metals while rejecting the alkaline-earth species.10 Over 60 metals can react with 8HQ to form complexes, with formation constants in solution ranging from 104 (Ba2+) to 1049 (Ga3+).10,11 Detailed knowledge of the chemistry between the interaction of metal ions with the immobilized 8HQ is critical to its successful application. Our understanding of the chemistry of metal ion binding of silica-immobilized 8HQ has generally relied on detection of solution-phase analyte breakthrough or elution from the end of a packed column.8,9,12-16 Jezorek and Freiser8 used these methods to report the first pH effects for metal ion complexation with 8HQ immobilized to controlled-pore glass (CPG). Marshall and Mottola9 derived a simple synthetic route to immobilize 8HQ onto CPG using a diazo-coupling reaction to a surface-bound aromatic amine; they determined the pKA of the surface-bound ligand and binding capacities for various metal ions. Cantwell and co-workers14-16 accounted for the calcium binding equilibria of immobilized 8HQ using a site binding model that included the effects of the silica surface potential. A pulsed elution technique was recently developed by Howard and Holcombe17 where a frontal chromatography model was applied to predict the behavior of immobilized 8HQ in flow systems. While column elution studies have produced important information about the binding capacity of immobilized 8HQ ligands, less progress has been made in developing in situ spectroscopic methods to detect the various forms of the ligand on the silica surface. Fluorescence spectroscopy has been adapted to flow (5) Hankins, M. G.; Hayashita, T.; Kasprzyk, S. P.; Bartsch, R. A. Anal. Chem. 1996, 68, 2811-2817. (6) Feng, X.; Fryxell, G. E.; Wang, L.-Q.; Kim, A. Y.; Liu, J.; Kemner, K. M. Science 1997, 276, 923-926. (7) Hill, J. M. J. Chromatogr. 1973, 76, 455-458. (8) Jezorek, J. R.; Freiser, H. Anal. Chem. 1979, 51, 366-373. (9) Marshall, M. A.; Mottola, H. A. Anal. Chem. 1983, 55, 2089-2093. (10) Marshall, M. A.; Mottola, H. A. Anal. Chem. 1985, 57, 729-733. (11) Stary´, J.; Zolotov, Y. A.; Petrukhin O. M. Critical Evaluation of Equilibrium Constants Involving 8-Hydroxyquinoline and its Metal Chelates; IUPAC Chemical Data Series No 24; Pergamon Press: Oxford, U.K., 1979. (12) Sturgeon, R. E.; Berman, S. S.; Willie, S. N.; Desaulniers, J. A. H. Anal. Chem. 1981, 53, 2337-2340. (13) Bernal, J. P.; Rodriguez De San Miguel, E.; Aguilar, J. C.; Salazar, G. De Gyves, J. Sep. Sci. Technol. 2000, 35, 1661-1679. (14) Vermeulen, D. M.; Cantwell, F. F. Anal. Chem. 1993, 65, 1360-1366. (15) Chow, P. Y. T.; Cantwell, F. F. Anal. Chem. 1988, 60, 1569-1573. (16) Vermeulen, D. M.; Cantwell, F. F. J. Chromatogr. A 1995, 693, 205-216. (17) Howard, M. E.; Holcombe, J. A. Anal. Chem. 2000, 72, 3927-3933. 10.1021/ac020252d CCC: $22.00
© 2002 American Chemical Society Published on Web 08/23/2002
injection methods for in situ detection of aluminum ion complexation to silica-immobilized 8HQ.18 In situ absorbance measurements conducted through a thin layer of immobilized 8HQ were used to determine the second acid dissociation constant of the ligand.19 In general, electronic absorption or fluorescence measurements of immobilized 8HQ cannot readily distinguish the various forms of the ligand. Vibrational spectroscopy provides the opportunity to distinguish these forms and to probe their structure. In recent work from our laboratory,20 we demonstrated that in situ fiber-optic Raman spectroscopy was capable of distinguishing the protonated, neutral, and metal ion-complexed forms of silicaimmobilized 8HQ. The purpose of the present work is to apply in situ Raman spectroscopy in an investigation of the interfacial chemistry of silica-immobilized 8HQ for binding of metal ions over a wide range of solution conditions. Since the derivatized silica has a high specific binding capacity (typically ∼0.2 mmol/g), the mass of silica to be equilibrated with solution needs to be small for studies of reactions with trace-level (µM) metal ions; otherwise, the solution volume required to reach equilibrium becomes excessive. For example, a 0.5-g sample of 8HQ-derivatized silica reacting with a 10 µM solution of a target metal ion would require 10 L of solution to reach equilibrium under strong binding conditions. To address this problem, a flow cell is designed for this work using a fiber-optic Raman probe inserted directly into the packed end of a microcolumn, allowing excitation and collection of Raman scattering from less than 10 mg of derivatized silica. This cell is attached to a flow system that allows control of solution conditions while the response of the 8HQ-silica material is acquired by continuous monitoring of Raman scattering from the sample. To account for the effects of changing surface potential on the protontransfer and metal ion binding reactions of the immobilized ligands, ζ-potential measurements are made on the 8HQ-silica particles under the same solution conditions that are employed in the Raman scattering measurements.
Figure 1. Structure of the silica-immobilized 8HQ, showing the two tautomeric forms of the ligand.
EXPERIMENTAL SECTION Synthesis of Surface-Immobilized 8HQ. The synthesis of the silica-immobilized 8HQ (see Figure 1) was carried out using the procedure developed by Marshall and Mottola9 using a diazonium salt of a silica-immobilized aromatic amine to immobilize the 8HQ. The specific reagents and conditions used for this synthesis were described in an earlier publication.20 The 60µm particle diameter, 60-Å-pore diameter silica gel (EM Science) was first derivatized with (p-aminophenyl)trimethoxysilane; elemental analysis (MHW Laboratories, Phoenix, AZ) of the silicabound aromatic amine yielded 1.54% C and 0.26% N, which corresponds to a surface coverage of 0.36 µmol/m2, based on the specific surface area (550 m2/g) of the silica gel. Subsequent reaction of the diazonium salt of the aromatic amine with 8HQ yields a final silica product containing 3.26% C and 0.69% N. Subtracting the mass of carbon and nitrogen of the original aromatic amine anchor groups from the final product values gives the increase in carbon and nitrogen derived from binding 8HQ to the surface; the carbon and nitrogen are both consistent with
an 8HQ specific capacity of 0.15 mmol/g or a surface coverage of 0.28 µmol/m2. This result indicates that ∼80% of the phenylamine anchor groups bind 8HQ in the final product. The surface coverage of the bound 8HQ is ∼1/10 of a full monolayer, corresponding to an average distance of ∼24 Å between ligands. Preparation of Buffers. The buffer system for this work needed to have spectroscopic transparency, a known and constant ionic strength, negligible fluorescence, and minimal complexation with copper ions in the pH 1-8 range. For these reasons, the buffers were prepared from hydrochloric acid (Fisher), chloroacetic acid (pKA ) 2.88 21), acetic acid (pKA ) 4.76 21), ACES (N(2-acetamido)-2-aminoethanesulfonate, pKA) 6.99 21, and phosphate, pKA2 ) 7.21 21), titrated with sodium chloroacetate, sodium acetate, and sodium hydroxide (Aldrich). Sodium chloride was added to yield a constant ionic strength of 0.1 M. The Cu2+ concentrations were sufficiently low so that they did not affect the ionic strength of the buffered solutions. The pH was of each solution was verified at 23 °C with a pH meter. ζ-Potential Measurements. A Phase Analysis Light-Scattering Zeta-Potential Analyzer (Brookhaven Instruments Corp.) was used to determine the ζ-potential of the silica-immobilized 8HQ at various solution pH values and Cu2+ concentrations. The ζ-potential is measured at the interface between the moving particle and the liquid known as the shear plane.22 This system measures the optical phase shift of the particles migrating under the influence of an electrical field of ∼8 V/cm. The 8HQderivatized silica gel used for these measurements was a LiChrosorb Si100 (EM Separations) chosen for smaller (5 µm) particle diameter to avoid settling during the light scattering measurements; for equivalent surfaces, the measured ζ-potential does not depend on particle size.22 Measurements were carried out at Cu2+ concentrations ranging from 0 M, 30 µM, 600 µM, to
(18) Weaver, M. R.; Harris, J. M Anal. Chem. 1989, 61, 1001-1010. (19) Kolstad, K.; Chow, P. Y. T.; Cantwell, F. F. Anal. Chem. 1988, 60, 15651569. (20) Uibel, R. H.; Harris, J. M. Appl. Spectrosc. 2000, 54, 1868-1875.
(21) Benyon, R. J.; Easterby, J. S. Buffer Solutions. The Basics; IRL Press: Oxford, 1996. (22) Hunter, R. J. Zeta Potential in Colloid Science: Principles and Applications; Academic Press: New York, 1981.
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002
5113
Figure 2. In situ Raman spectroscopy instrument used to monitor metal ion complexation to silica-immobilized 8HQ. See text for description.
5 mM, and solution pH was varied from 2.3 to 9.8 to cover the range of proton-transfer and metal ion binding equilibria. The ionic strength was held constant at 0.1 M so as to not change the Debye length of the double layer. An average of six runs was reported for each measurement. Raman Measurements. The Raman instrument (shown in Figure 2A) was previously described;20 briefly, a 647.1-nm line from a krypton ion laser (Coherent, Innova 90-K) was passed through a dove prism, a plano-convex lens (f ) 1.0 m), and a 0.5-mm pinhole to remove plasma lines. The beam was focused onto a bifurcated fiber-optic probe (Fiberguide Industries) using a multimode fiber coupler (Newport) and a 10× microscope objective. The bifurcated fiber-optic probe consisted of 44 silica collection fibers (100 µm) surrounding a central excitation fiber. The light collection fibers were arranged in a linear array, collimated by an F/1.3 camera lens (JML Optics) and focused by a plano-convex lens (f ) 175 mm) onto the entrance slit of a F/7 single-stage 0.5-m spectrograph (Spex). A holographic notch filter (Kaiser) for the 647.1-nm line was placed between the collection and focusing lenses to remove Rayleigh and specular scattering. The Raman scattering was detected with a 1024 × 256 TE-cooled CCD (Andor). Samples were typically excited with ∼100 mW of incident laser power, and scattering was integrated in six accumulations of 20 s each. The microflow cell used in the metal ion binding studies was a modified stainless steel 1/4-in. Swagelock Tee shown in Figure 2B. The cell was assembled by inserting the common end of the fiber optic into one side of the Tee and then inserting a polished 5114
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002
0.075-in.-inner diameter tubing into the other side of the Tee. The fibers and small-diameter tubing were pressed together, and connectors on the ends of the Tee held these parts in place. The sidearm of the Tee was used as the solution outlet. The smalldiameter tubing was then filled with ∼10 mg of dry surfacemodified silica particles and connected to the flow system. An acetate buffer solution (pH 4.7) flowed through the column at a rate of 2 mL/min with the aid of an HPLC pump (Beckman, model 110A) to pack the dry silica up against the fiber optic. Changes in the 8HQ Raman signal were used to monitor the packing of the silica, and typically, times for the Raman signal to reach steady state were ∼10 min. The acetate buffer was pumped at 2 mL/ min for another 30 min to ensure that all of the silica particles had packed as closely as possible. To equilibrate metal ioncontaining samples with the silica surface, typically 100 mL of buffered solution was pumped through the column at 1.0 mL/ min. To remove metal ions bound to the surface, aqueous 0.1 M HCl was used as an eluent to reset the surface for subsequent experiments. Data Analysis. Prior to data analysis, an observed Raman spectrum was baseline-subtracted by fitting the fluorescence background to a cubic equation using 10 points that were varied until the difference spectrum contained a flat baseline with zero offset; the resulting difference spectrum was then normalized. Classical least-squares regression analysis was employed to test complexation models and to resolve the spectra of the different ligand components. Raman spectra were collected during a series of different runs at constant metal ion concentrations and variable pH. These spectra were collected into the columns of the data matrix, D, which contains r rows, which indicate the spectral wavenumber dimension, and c columns, which denote the different samples at varying pH. The data matrix can be expressed as the product of the spectral and concentration behaviors of the n individual components in the sample:
D ) AC
(1)
where A is an r × n matrix of n pure component Raman spectra and C is an n × c matrix that describes the concentration behavior of each component versus pH. The goal of this study is centered on investigating changes in concentrations of ligand forms on the silica surface; the first step of this analysis is to estimate the number of varying components that are in the data set. Malinowski23,24 developed a test to determine the number of significant components during a run by applying a variance ratio or F test25 to the reduced eigenvalues (REV) of the data matrix, D. The reduced eigenvalues are proportional to the amount of variance in the data that is described by their corresponding eigenvectors. The F ratios are obtained by a two-way test between the first REV and the second REV, and each subsequent element of F is the two-way test between the subsequent REV and the one after it. A large F ratio indicates that the addition of a component to the model describes significant variance in the data; the F ratios are tested for significance at the 95% confidence level to determine the number of independently varying components in the data. (23) Malinowski, E. R. Factor Analysis in Chemistry, 2nd ed.; John Wiley and Sons: New York, 1991. (24) Malinowski, E. R. Anal. Chem. 1977, 49, 612-617. (25) Barlow, R. J. Statistics a Guide to the Use of Statistical Methods in the Physical Sciences, John Wiley and Sons: New York, 1989.
Once the number of independently varying components is determined, the data matrix can be tested for the presence of component spectra of the surface ligand, determined in advance in the absence of metal ion or in metal ion excess, and at specific regions of pH (see below). Using these predetermined pure component spectra, A, a least-squares estimate of the corresponding concentration profile, Cˆ , can be found by left-multiplying the data matrix D by the pseudoinverse of the A matrix:26,27
C ˆ ) [AAT]-1ATD
(2)
which was evaluated using Matlab 4.2. The concentration variation of the components is then fit to a site binding model, C, where the algebraic solution is evaluated using Maple version 5. The parameters in the model are varied to minimize χ2, χ2 ) (1/σ2)∑R2, where σ2 is the variance of the resolved composition profiles and ∑R2 is the sum of squares of the residuals, R ) C ˆ - C. The estimated parameter uncertainty for the optimized site binding model were calculated by mapping the χ2 surface within a 15% range the best fit of the parameter, p, and fitting the error surface to a parabola to determine its curvature. The variance in the parameter thus determined is σp2 ) 2/(δχ2/δp2).28,29 Replicate measurements resulted in optimized parameter values that were found to lie within the uncertainty estimated from the error surface curvature. RESULTS AND DISCUSSION Raman Spectra of Silica-Immobilized 8HQ. Our first experiments in this study were to establish whether it would be possible to detect and distinguish the different chemical forms of the silica-immobilized 8HQ ligand in a 10-mg sample of derivatized silica packed in the microvolume flow cell. From the surface coverages of the derivatized silica (see above), a 10-mg sample corresponds to 1.5 µmol of the immobilized 8HQ in the flow cell. A small amount of ligand is necessary so that equilibrium can be achieved with reasonable volumes of solutions containing low concentrations of metal ion, typical of those found in trace-level metal preconcentration and analysis. The equilibration of surface ligands with low concentrations of Cu2+ ions in solution could be achieved with reasonable solution volumes. For example, 100 mL of a 30 µM Cu2+ solution (100% excess), loaded in 100 min at 1 mL/min, was more than sufficient to equilibrate the immobilized ligands, as evidenced by the Raman spectrum reaching steady state after ∼70 mL of metal ion solution passed through the flow cell. The spectra of the deprotonated, protonated, neutral, and copper-complexed forms of the silica-immoblized ligand were acquired and are presented in Figure 3. The deprotonated (anionic) form of the ligand was generated by equilibrating the sample with pH 10.6 phosphate buffer. The protonated form of the ligand was generated by rinsing the column with 0.10 M HCl. The neutral form of the ligand was formed by flushing the column with acetate buffer, pH 4.7; finally, the copper complex with the (26) Draper, N. R.; Smith, H. Applied Regression Analysis; John Wiley and Sons: New York, 1981. (27) Tauler, R.; Kowalski, B.; Fleming, S. Anal. Chem. 1993, 65, 2040-2047. (28) Bevington, P. R. Data Reduction and Error Analysis for the Physical Sciences; McGraw-Hill Book Co.; New York, 1969. (29) Phillips, G. R.; Eyring, E. M. Anal. Chem. 1988, 60, 738-741.
Figure 3. Normalized Raman spectra of silica-immobilized 8HQ: deprotonated (top), protonated (second down), neutral (third down), and complexed to Cu2+ (bottom).
immobilized ligand was generated at pH 4.7 by equilibrating the sample with 10 mM Cu2+. It is clear from the resulting spectra that the protonation and metal ion binding states of the ligand can be distinguished in the in situ Raman spectra. Much of the differences between these spectra arise from changes in the fractions of tautomeric forms of the ligand (see Figure 1) in response to pH and metal ion binding.20 The symmetric sNd Ns stretching vibrations of the azo group are characterized by a strong Raman scattering in the region 1400-1450 cm-1 due to νNdN;30 this band is observed at 1431 cm-1 for the anion and copper complex form of the ligand. The two bands at 1141 and 1196 cm-1 for the anion and complexed ligand correspond to the C-N azo symmetric stretch, and C-N azo symmetric bend, respectively. The frequencies of these three bands are identical to the uncomplexed azo-quinoline species30 which indicates that the metal ion coordination does not involve N-atoms in the azo bridge but the stronger N,O sites of the quinoline ring. The 1141- and 1431-cm-1 C-N azo bands are also observed in the neutral form of the ligand at smaller intensities when compared to the anion and complexed form. The protonated form has an even more diminished 1141-cm-1 band, and its νNdN band at 1431 cm-1 appears only as a shoulder. The intense peak at 1378 cm-1 is strong in the metal ion complex form of the ligand; this band is due to the νC-C mode of the quinoline ring31 and appears as a shoulder in the neutral and deprotonated forms of the ligand. The decrease in the intensities of both of the azo and quinoline bands for the protonated and neutral forms of the ligand is (30) Trotter, P. J. Appl. Spectrosc. 1977, 31, 30-35. (31) Bajpai, P. K.; Pal, B.; Baul, T. S. B. J. Raman Spectrosc. 1995 26, 351-361.
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002
5115
consistent with formation of a hydrazone tautomer.30,31 The large band at 1590 cm-1 corresponds to the νCdC stretching vibrations of the phenol ring, and the peak intensity of this band for the uncomplexed ligand forms is strong, consistent with hydrazone formation.31 Further evidence of the presence of this tautomer is the shoulder at ∼1620 cm-1, which has been assigned to νCdN and νCdO in hydrazones of azonaphthols and azoquinolines.30 Other evidence of this tautomer for both the protonated and neutral ligands is the 1283-cm-1 νC-C in-plane stretching band that is associated with the hydrazone form of the ligand. The observed spectrum shows that the protonated ligand is predominately in the hydrazone form, the neutral and anion ligand exists in both tautomers (the anion exhibiting more azo-quinoline character), and the metal-complexed ligands are predominately in the azo form. Deprotonation at high pH drives the ligand toward the azo form to favor phenoxide anion formation on the more electronegative oxygen over an azide ion by loss of hydrogen ion from the nitrogen linker portion of the molecule. The spectra of the metal ion complex are consistent with metal ion binding to the phenoxide anion formed from the azo tautomer of the ligand, which also allows interactions with the nitrogen in the quinoline ring. Interfacial Metal Ion Complexation and Surface Potential. Free-solution complexation of 8HQ with metal ions generally exhibits a stoichiometry that is the same as the metal ion charge.11 For example, Cu2+ forms a binary 8HQ complex in free solution with formation constants 1012 and 1011 for binding the first and second ligands, respectively.11,32 One difference between solutionphase and silica-bound ligands is their ability to form higher-order complexes. At low surface coverages, immobilized ligands may be spaced too far apart to allow formation of higher-order complexes, so that the complexation ratio of metals to silicaimmobilized 8HQ is limited to 1:1.20 A second difference in interfacial metal ion complexation is that the reaction takes place in a different chemical environment than bulk solution due to the surface potential of the silica, which will alter the chemical potential of ions in the vicinity of the surface.14-19 The silica gel itself exhibits a surface charge due to deprotonation of silanols, and nonspecific adsorption of other ions from solution to the silica surface can alter the potential. The surface potential may affect the reactivity of a surface ligand since ion activities near the surface will be altered from their bulk solution values due to the surface potential. Since the binding site of immobilized 8HQ is tethered by a relatively stiff linkage well away from the silica surface, the complexation reaction takes place beyond the Stern layer of strongly adsorbed ions in the diffuse layer of solution phase ions; the ion densities present in this region can be modeled by Gouy-Chapman theory.33 The relationship between the activity of cations at a distance x away from the charged surface (aj i,x) and in bulk solution (ai) is given by the Boltzmann equation:
aj i,x ) ai exp(-zFψx/RT)
(3)
Figure 4. Zeta potential of the silica-immobilized 8HQ. Full plot are results without Cu2+ in solution. The inset plot shows the zeta potential of solutions containing the Cu2+ at 5mM (top), 600 µM (middle), and 30 µM (bottom).
surface, z is the charge on the ion, F is Faraday’s constant, R is the gas constant, and T is the absolute temperature of the solution. The Boltzmann relationship is only strictly true for ion densities expressed as activities; however, the relationship is routinely applied to concentrations when the concentrations are small.33 The binding of Cu2+ due to complexation by the immobilized 8HQ thus depends on the electrical potential of the silica surface and its effect on cation exchange in the diffuse part of the electrical double layer. The surface potential of the silica arises primarily from protonation and deprotonation of silanol groups34 and from the adsorption and binding of metal ions at the silica surface.14,16,18,22 To understand and model the effect of surface potential on interfacial complexation reactions by immoblized 8HQ, we measured the ζ-potential of the derivatized silica for several different pH solutions containing three different Cu2+ concentrations to span the range of spectroscopic studies. The ζ-potential of the 8HQ-silica for each of the solution conditions is plotted in Figure 4. When no Cu2+ is present in solution, the ζ-potential is negative for all solutions above a pH of 2.7 due to deprotonation of silanol groups; at a pH
